Question

#include <iostream>
#include <vector>

int main()
{
    std::vector<double> test1(1);

    test1[100] = 5.;

    std::vector< std::vector <double > > test2(1, std::vector<double>(1,0));

    test2[50][100] = 5.;
}

test1 : resizes and allocates memory nicely

test2 :"Segmentation fault (core dumped)". Why?

Note: Cannot use matrix, since row size are unequal.

Summary:
at(int): checks bounds and throws exception if necessary – no resizing

operator[](int) :doesnot check bounds – no resizing

push_back() : resizes increases capacity() by double if current capacity is small

size() : number of elements in vector

capacity() : maximum elements to hold before a re allocation is necessary

Answer 1

You access element with index 100 of a vector of size 1. You should not access index out of a vector bounds. Truth is that it is pure luck that it works in the first case instead of strange that second does not work.

A vector expands on call to resize() or push_back, but simply accessing an index does not expand the vector’s size. Instead it causes an undefined behavior.

To fix the code do(change the sizes used when constructing the vectors):

#include <iostream>
#include <vector>

int main()
{
    std::vector<double> test1(101);

    test1[100] = 5.;

    std::vector< std::vector <double > > test2(51, std::vector<double>(101,0));

    test2[50][100] = 5.;
}